4.1 Summary/Abstract The Visualization TRD addresses an ongoing demand for software that allows biomedical scientists to visualize their simulation and experimental data, which is often large-scale. Furthermore, it will produce a set of uncertainty visualization tools to help scientists directly explore simulation and measurement results, design and troubleshoot experiments, and to help better understand the relationship of multiple parameter uncertainties to simulation results. This project leverages world-renowned research in visualization from the Scienti?c Computing and Imaging (SCI) Institute to address the timely challenges of visualizing large datasets and the increased need to visualize associated uncertainty in simulation and measurements. The proposed software tools and methods will address the problem of understanding large biomedical volumes with our Driving Biomedical Projects through the devel- opment of better algorithms and software for manipulating, rendering, and characterizing uncertainty in large, complex volume datasets, and by working closely with the TRD Image and Geometric Analysis, the development and integration of new types of algorithms for the analysis of large volumes. These new algorithms will be de- veloped in concert with an evolving set of software tools that are designed to address the speci?c needs of our Driving Biomedical Projects and a set of collaborating scienti?c investigators and they will be actively dissemi- nated to the community at large. Speci?c Aims: Methods/infrastructure for the visualization of large volume datasets This aim entails the development of technologies that will enable interactive, 3D visualization of both large volumes (3D images) of data and large databases of volumetric data. To achieve this progress, we will develop new approaches to creating as well as rendering levels of detail, streaming methods for large data. Uncertainty visualization This aim entails novel approaches for the quanti?cation and visualization of uncer- tainty in image analysis and simulation problems with complex and variable geometries and uncertain parameters, overcoming occlusion and clutter that obscure important regions of interest. Visual analysis of large volumetric datasets This aim addresses the development of technologies and software for analysis of large volumetric datasets to support ef?cient, effective visualization. To this end, the investigators will develop new technologies for creating aggregations and summaries, and user-guided feature identi?cation and tools that interact seamlessly with quantitative analysis. These developments will be speci?cally designed to address the challenges of interactive visualization of large volumetric datasets, as in Aim 1.